The Photog Plan
Overview
The Photog Plan is a web application I built for wedding photographers as a full-stack learning project. I designed it, built it, and operate it solo — from provisioning the bare VPS to handling end-user support. It’s free and donation-based, running on infrastructure I administer myself. I keep it live as a foundation I can host future projects on top of.
Infrastructure
The application runs on a Linux VPS that I provisioned and harden. The full stack:
- Web server: nginx — handles TLS termination, reverse proxying, static file serving, and access/error logging
- DNS: records managed directly for the domain
- TLS: certificates provisioned and renewed for HTTPS
- Firewall: configured to expose only necessary ports
- SSH: key-only authentication, password login disabled
- Patching: I own the OS and package update cycle for the server
- Backups: configured and monitored
- Logging: nginx access and error logs for monitoring and troubleshooting
Application Stack
- Frontend: HTML and JavaScript
- Backend: Node.js
- Database: Supabase (PostgreSQL) with row-level security policies that enforce data isolation between users
Row-level security is the critical piece: it ensures that one user cannot access another user's data at the database level, regardless of what the application code does. Even if there were an application-layer bug, the database enforces tenant isolation.
Security Design
I built this application with an attacker's perspective. Every security boundary was designed against the flaw classes I hunt in my bug bounty and penetration testing work:
- Authentication: designed to resist credential stuffing, session fixation, and token leakage
- Session handling: secure token management with proper expiration and invalidation
- Access control: row-level security at the database layer, not just application-layer checks
- Input validation: server-side validation against injection and XSS vectors
Building something you intend to attack-test yourself changes how you write every line. The threat model isn't theoretical — it's the same checklist I work through on bug bounty targets.
Operations
Running a production web application means owning everything after the code ships:
- Server uptime monitoring and failure diagnosis
- OS and dependency patching on a live system
- Backup verification
- Log review and incident response
- End-user support — translating customer requirements into technical work
This is the same operational responsibility that comes with running infrastructure in any environment. The difference is that I do it entirely on my own, so every decision about availability, security, and maintenance is mine to make and mine to live with.
What This Demonstrates
- Full-stack ownership: from bare server provisioning to production application to end-user support
- Infrastructure hardening: key-only SSH, firewall configuration, TLS, patching discipline
- Security-first development: applying offensive knowledge to defensive design decisions
- Operational maturity: maintaining a live, internet-facing service with real users